Research on nonlinear filtering technology has been conducted to address the challenges faced by unmanned ships, which undergo significant swaying motion and are subjected to complex external interference under harsh sea conditions. Firstly, to tackle the issue of the measurement noise variance matrix not being constant and not meeting Gaussian characteristics under single and short-term strong interference, an optimized UKF algorithm was developed. This algorithm draws on the application of sliding windows and the advantages of multi-channel adjustment in two filtering algorithms. The optimized UKF algorithm uses distance judgment to re-weight the innovation vectors within the sliding window, estimates the measurement noise variance matrix, and establishes an adaptive factor matrix to adjust the measurement noise variance matrix. Additionally, by introducing an anomaly judgment factor twice, the algorithm ensures alignment accuracy while significantly improving alignment speed, as verified through simulation.